The Great Escape? Low pay and progression in the UK s labour market

Transcription

1 The Great Escape? Low pay and progression in the UK s labour market October 2017 Conor D Arcy and David Finch Resolution Foundation Social Mobility Commission Sanctuary Buildings 20 Great Smith Street London SW1P 3BT contact@smcpcommission.gsi.gov.uk 1

2 About the Commission The Social Mobility Commission is an advisory non-departmental public body established under the Life Chances Act 2010 as modified by the Welfare Reform and Work Act It has a duty to assess progress in improving social mobility in the UK and to promote social mobility in England. It currently consists of four commissioners and is supported by a small secretariat. The Commission board comprises: The Rt. Hon. Alan Milburn (Chair). The Rt. Hon. Baroness Gillian Shephard (Deputy Chair). Paul Gregg, Professor of Economic and Social Policy, University of Bath. David Johnston, Chief Executive of the Social Mobility Foundation. The functions of the Commission include: Monitoring progress on improving social mobility. Providing published advice to ministers on matters relating to social mobility. Undertaking social mobility advocacy. This work contains statistical data from ONS which is Crown Copyright. The use of the ONS statistical data in this work does not imply the endorsement of the ONS in relation to the interpretation or analysis of the statistical data. This work uses research datasets which may not exactly reproduce National Statistics aggregates. 2

3 Contents Executive Summary... 1 Section 1: Introduction... 4 Progression and policy... 4 Definitions and methodology... 5 Section 2: Who escapes, cycles and gets stuck... 7 Section 3: What helps people escape low pay? Section 4: Conclusion Annex

4 4

5 Executive Summary The coming years will be pivotal for the UK s labour market. Along with new questions raised by Brexit, the responses to longstanding issues dogging firms and workers alike will be key. When it comes to action points, Britain s woeful productivity growth since the financial crisis is top of the list but not far below is getting to grips with a decades-long overreliance on low-paid jobs. Sometimes overlooked is the extent to which these two issues are connected: a more productive economy is likely to be built upon a more highly-skilled, highly-paid workforce. Answering the question of whether low pay is the first step towards betterpaid, better-skilled roles or as high as many employees climb is therefore an essential prerequisite in responding to the productivity challenge, as well as making sure the country works for all earners. To explore this, our analysis tracks initially low-paid employees (earning below twothirds of median hourly pay) across a decade. Based on their outcomes, we split them into three main groups: Stuck : employees who were in low-paid work in every year they appear in the data. Escapers : those who earn above the low pay threshold in each of the final three years of the decade, suggesting they have made a sustained move onto higher wages. Cyclers : people who fall between the above categories, moving onto higher wages at some point during the decade but not consistently out of low pay by the end of the period. Our analysis finds that for most low-paid workers, poorly-paid positions are not acting as a first rung on the ladder it is the only rung. Of all those low paid in 2006, by 2016 just one in six (17 per cent) were escapers. One in four (25 per cent) remained stuck throughout the period while just under half (48 per cent) were cyclers, moving onto higher wages at some point but not sustaining that progress. The remaining one in ten employees exited the data, meaning they were not an employee after the initial period. Despite this negative overall picture, these figures represent progress over recent decades. For instance, 35 per cent of those who were low paid in 1981 were still stuck in 1991, the highest proportion in our data which stretches back to low paid workers in But while the share of low-paid employees escaping has risen slightly over the past quarter of century (from one in ten, or 11 per cent, for the cohort), the falling share of people becoming stuck appears to have been replaced by a greater proportion of cyclers. Change and some progress has been visible within groups too. We find that women are overrepresented among those who remained stuck from 2006 to

6 But the proportion of women getting stuck has fallen from 48 per cent in to 30 per cent in , when those who exit the data over the following decade are excluded. In contrast, the risk of long-term low pay has increased for men over the same period (from 20 per cent to 25 per cent). This is likely due to the increasing number of men working in low-paid part-time work. People switching from part-time to full-time work were more likely to have escaped than those who didn t. In order to assess which of these characteristics appear to be linked to getting stuck or escaping, and the relative size of the effect those factors have, we carry out regression analysis to isolate the key factors. We find that the number of years spent in employment over the following decade is positively and significantly linked to escaping from low pay. On average, 2016 escapers had been in employment for an average of 8.3 years compared to 5.9 years for cyclers and 4.5 years for the stuck. This is likely to weigh particularly heavily on women who have children and take time out of the labour market after giving birth. Our analysis finds that women are significantly less likely to escape than men, with the size of the negative effect particularly large for women in their early 20s. This, combined with the fact that the number of years worked part-time is also negatively linked to progression strongly suggest that caring responsibilities and the lack of quality part-time work is bearing down on the progression chances of mothers. This reinforces previous research on the topic, and organisations such as Timewise have long highlighted the dearth of quality employment available for women with caring responsibilities. But given the scale of the skills and potential of mothers that is currently going untapped, more clearly should be done. Some of this responsibility lies with employers, ensuring where possible that they offer more higher-paying roles on a part-time or flexible basis. Government has a key role too, with the design of Universal Credit and the incentives it offers to women to seek out better-paying work currently doing too little to support progression. And while the new system of in work conditionality will most likely push affected claimants towards full-time work rather than higher-paying positions, government, as well as local leaders like Metro Mayors, should take this shift in the role of the state as an opportunity and be thinking creatively about what interventions could help more low-paid women to progress. Some kinds of workplace are also significantly linked to the question of progression. Accounting for the type of occupation worked in, a strong negative impact persists for those who remain in the accommodation and food services industry. This is unsurprising in some senses given the prevalence of low-paying roles in the sector but does suggest that people find it easier to progress by leaving hospitality and moving into other higher-paying industries. But the issues already highlighted on Brexit and productivity are particularly relevant for hospitality given its relatively large EU workforce and low productivity compared to other countries. Taking progression seriously and building more routes out of low-paying roles and into higher-skilled positions within the industry would be an important step. 2

7 The government s forthcoming industrial strategy could contribute to this by putting not only high-employment (in which the UK has some considerable success) but reducing the prevalence of low pay in sectors in which it is endemic, such as hospitality and retail, front and centre. Focusing on these parts of the economy will be essential if low pay is to become a springboard to higher earnings for more people and to improve our long run record of only a minority of low paid workers escaping to a higher rate of pay within a decade. 3

8 Section 1 Introduction Progression and policy Low pay is a longstanding problem for the UK, with the country regularly appearing near the bottom of international league tables. The proportion of people who are classed as low paid has moved little in 20 years, with approximately one in five employees meeting this definition. But this snapshot, even repeated over decades, only tells us so much about the dynamics within low pay. Pay is a key determinate of living standards. And while it is not true that all low-paid workers live within low-income households, persistent low pay certainly presents a challenge to many families trying to make ends meet. If low pay is difficult to escape from, the policies that follow are likely to be different from those in a world in which low-paid jobs are a springboard to higher earnings. Similarly, if certain kinds of workers appear to face higher barriers to finding better-paid work, targeted responses may be needed to support them. The question of policy is particularly relevant at this moment. The UK s productivity woes since the financial crisis are well-known, but the link to the country s overdependence on low-paid work is often glossed over. Given research that shows the gap between the UK and other large advanced economies is driven by productivity deficiencies in most sectors, 1 including low-paying ones, taking low pay and progression seriously should be part of any plan to raise productivity. One policy already in place that could improve matters is the National Living Wage (NLW). As recent Resolution Foundation research has illustrated, the NLW has already begun to reshape the lower end of the labour market, with minimum wage earners receiving pay rises well ahead of average pay growth. 2 While this is a welcome contrast to the falling real wages experienced by many in 2017, it brings new potential challenges with around 4 million people likely to be paid at or close to the wage floor by What this will mean for routes out of low pay should be considered now, rather than only when the NLW reaches its target at the end of the decade. The continued rollout of Universal Credit is another piece of the policy puzzle around progression. While most of the public debate has understandably focused on the most immediate impacts of the new system such as the difficulty a six-week wait for payment presents for many claimants in-work conditionality marks a significant shift from existing practice. With exceptions for those with young children, caring responsibilities or ill health, people earning less than the equivalent of 35 hours a week at the wage floor will be required to either increase the number of hours they 1 ONS, International comparisons of labour productivity by industry: 2014, October C D Arcy, Low Pay Britain 2017, Resolution Foundation, October

9 work or the rate they earn. While many may move into full-time work at their same hourly rate, this new approach could potentially mean more people especially lowpaid mothers are seeking better-paying opportunities. And ultimately policy has the potential to do far more to support pay progression beyond a full-time job on the minimum wage. The question of pay progression interlocks with concerns around the intergenerational impact of various living standards trends. As the Resolution Foundation s Intergenerational Commission has explored, younger people have experienced a deeper dip in their earnings than their older colleagues since People in their 20s remain much less likely to switch jobs a key catalyst to pay growth especially for younger workers than in the past and those remaining in the same job for longer are not getting a bigger pay lift. The lack of robust pay growth has short-term consequences of course but is also like to delay or derail aspirations like buying a home or saving for retirement. For young people at a key point in their career trajectory, this could mean permanent scarring. The extent to which young people are escaping from low pay therefore has important consequences for today and tomorrow with an equal need for policies that can ameliorate the situation in both the short and long term. Definitions and methodology To help inform the debate around progression, this report explores the earnings trajectories of low-paid people and how that has varied over recent decades. We then assess how the likelihood of moving onto higher wages differs across employees and identifies the factors that appear linked to progression or remaining in low pay. In order to monitor these trends, explicit definitions of low pay and what it means to progress are required. Although a variety of definitions of low pay and progression have been utilised in other analysis, this report returns to those used in previous Resolution Foundation research for the Social Mobility Commission. 3 Low pay is defined as hourly earnings below two-thirds of the median hourly wage, excluding tips, commissions or other premium payments. For context, the low pay threshold taking this approach was 8.07 in 2016 the most recent data used in our analysis the minimum wage in the same period was 7.20, and the median was Our analysis primarily takes all low-paid people in a given year and tracks their pay over the following decade. We then divide them into four groups depending on their experience over the period: Stuck : people who were in low-paid work in every year they appear in the data. 3 C D Arcy and A Hurrell, Escape Plan: Understanding who progresses from low pay and who gets stuck, Social Mobility Commission/Resolution Foundation, November

10 Escapers : those who earn above the low pay threshold in each of the final three years of the decade, suggesting they have made a sustained move onto higher wages. Cyclers : people who fall between the above categories, moving onto higher wages at some point during the decade but who are not consistently out of low pay by the end of the period. Exiters : people who are no longer present in the data at the end of the period. Because the data relates solely to information about employees held by employers we cannot pin down why any specific person leaves employment but they may move into self-employment, retirement or other forms of inactivity, unemployment, leave the country or die. The dataset we use is the ONS s New Earnings Survey Panel Dataset (NESPD), the panel version of the Annual Survey of Hours and Earnings (ASHE) and its predecessor the New Earnings Survey, a large survey of employers about a sample of their employees. This means the pay data is likely to be more accurate than selfreported alternatives; many people do not know exactly what they earn per hour. It does however include less information on personal characteristics, some of which like education, training and attitudes are valuable in understanding why some people progress and others don t. 6

11 Section 2 Who escapes, cycles and gets stuck The detailed data stretching back to the 1970s allows us to explore how trends in terms of the share of low paid people that escape, and who escapes, has varied over time. In particular, we compare the two most recent decades available in the NESPD: and Before splitting the low-paid population into the groups defined above, we explore how the hourly wages of those who were low paid in the past grew over the following decade of their careers. Figure 1 begins by focusing just on those who were low paid in The median or typical employee in low pay in 2006 and also in work ten years later earned 8.75 per hour in 2016, above the low pay threshold of Figure 1: Distribution of pay progression for low paid workers in Source: Resolution Foundation analysis of ONS, NESPD But Figure 1 illustrates the wide range of outcomes for this group. Those at the 90 th percentile paid more than 90 per cent of the group were paid in the latest data. The lowest earners on the other hand at the 10 th percentile were paid 7.20 an hour, the level of the NLW in April Of course, those who were low paid are likely to have been a diverse group. The low paid in 2006 are likely to have included university students in low-paid work during their studies as well as those with much 4 The median here is above the low pay threshold, despite more than half of the 2006 low paid remaining stuck. This is because not all of those who are stuck (and so in low pay in the final three years of the period, explained in more detail below) are in work in the final year of the decade. 7

12 lower qualification levels; differing outcomes are therefore to be expected. But it highlights that for many, wage gains since have been modest. How does this progress compare with that made by low-paid groups in the past? Figure 2 contrasts the wage gains experienced by those who were low paid in 1996 over the following decade with the 2006 cohort. The more recent group have experienced weaker wage growth across the distribution. However, the gap is by far the largest for the lowest earners. While the wages of the 10 th percentile in 1996 rose by 41 per cent over the following decade, for the same group in 2006 the comparable figure was just 12 per cent. This is perhaps to be expected given the importance of the minimum wage for these very low earners. While the 1996 cohort were beneficiaries of the original introduction of the National Minimum Wage (NMW) in 1999 and successive aboveinflation increases, those low paid in 2006 worked through many years during which the NMW fell in real-terms. But despite the wider disparity at the bottom, the impact of the wage squeeze and the historically terrible decade the UK has experienced in pay terms is visible across the distribution. Figure 2: Decadal pay progression for low paid in & , additional to average pay progression of all employees (CPI-adjusted) Source: Resolution Foundation analysis of ONS, NESPD 8

13 We next turn to how our different groups escapers, cyclers, the stuck and exiters have changed over time, with Figure 3 showing variation over the past forty years. While patterns have varied substantially over time, there are some identifiable trends in the earnings progress of these groups. It is clear that most people do not escape from low pay. At all periods, the share of low-paid people in a given year that go on to escape low pay is the smallest of our three core categories (excluding exiters). That said, Figure 3 does suggest some improvements have been made in recent years, with 19 per cent of those low paid in 2003 having escaped by 2013, well above the 11 per cent escaping in the earlier years of the data. The most recent years has seen that progress slide somewhat, with one in six (17 per cent) of the low paid in 2006 having escaped by Figure 3: Proportion of low-paid workers by low-paid status ten years later Source: Resolution Foundation analysis of ONS, NESPD The largest group throughout the four decades are the cyclers. Although their precise share has varied over time, a trend over the past twenty years or so is a rise in the proportion of cyclers 48 per cent of those low paid in The majority of this increase appears to have come from a smaller number of people becoming stuck. From a high point of 35 per cent of those who were low paid in 1991 still being low paid in 2001, the proportion has now fallen to 25 per cent for those who were low paid in The number of people that exit the data has fallen over time to roughly one in ten (11 per cent) of the low paid in The pay trends discussed in Figure 1 map on to these groups too. As would be expected, the stuck experience the least earnings growth. As Figure 4 shows, in 9

14 2006 those who would go on to become stuck earned 6.75 in 2016 terms, just a little less than the future cyclers ( 6.90) and escapers ( 6.95). But across the following decade, the stuck barely made pay progress and earned, on average, 6.90 over the 10 years. By 2016, for those still in work who were stuck, the progress amounted to just 40p per hour wage growth across the entire period. At the other end of the spectrum, the escapers averaged across the period still less than the average wage of those who weren t low paid in 2006 across the decade ( 18.15) though this does not take age or other potentially important factors into account. The cyclers, interestingly, appear more similar to the escapers than the stuck. Their average pay over the period was 9.40 suggesting that genuine earnings progression has been made for many in this group at some point but without that being sustained. Figure 4: Average hourly pay by low-paid status, (CPI-adjusted) Source: Resolution Foundation analysis of ONS, NESPD We next turn to the composition of each of these groups. Women are more likely to be low paid, though men have faced an increasing risk of low pay in recent decades; in 1975, just 28 per cent of men were low paid. By 2006, this had risen to 36 per cent with women the remaining 64 per cent of the low paid. Figure 5 explores what proportion of men and women escape, cycle and get stuck. Although the differences are far from vast, women are more likely to become stuck in low pay, with 30 per cent of those who were low paid in 2006 still stuck by 2016 whereas the corresponding figure for men was 25 per cent. Men s likelihood of cycling and escaping low pay was slightly higher than for women. 10

15 Despite this more negative overall picture for women, the gains made over the past 25 years have been stark. Almost half of low-paid women (48 per cent) in 1981 were still stuck by While this 18 percentage point increase has been split between a rise in the share of cyclers as well as escapers, it nonetheless suggests that some progress has been made by women. Figure 5: Stuck, cyclers and escapers by sex, Notes: Totals exclude those who exited the data over the sample period. Source: Resolution Foundation analysis of ONS, NESPD A crucial factor in pay progression is age. Previous Resolution Foundation research for the Social Mobility Commission showed that, in general, pay rises more quickly while people are in their twenties with the upward progress tapering off sooner for non-graduates than for graduates. 5 But to what extent does low pay become stickier as people age? Figure 6 compares the difference between the proportion stuck, cycling and escaping among each age group, and the average rate across all ages for the period from 2006 to The greatest percentage point differences from the overall average are among the youngest and oldest groups. Those who were under 25 in 2006 were by far the least likely to be stuck in 2016 with just 15 per cent getting stuck, 13 percentage points lower than the overall average (again excluding those 5 C D Arcy and D Finch, Finding your routes: Non-graduate pathways in the UK s labour market, Social Mobilitiy Commission/Resolution Foundation, May

16 that exit). This lower likelihood is likely to be driven both by students and other people in education working in low-paid positions as well as the odds of moving up over time being higher when young, as well as the reduced odds of escaping as one ages a topic later analysis will return to. At the other end of the age range, older workers are overrepresented among the stuck. Of those aged 56 and over in 2006, 57 per cent were still stuck in Figure 6: Low pay status in 2016 by age in 2006 Percentage point difference from overall average Source: Resolution Foundation analysis of ONS, NESPD Another important piece of the progression puzzle is the region or nation in which the low paid live: is it easier to progress in generally higher-paying parts of Britain? Figure 7 illustrates that, compared to age, the differences between each region or nation and the British average are small. There are some notable exceptions however. The most obvious of these is Scotland. The share of low earners escaping low pay in Scotland was 4.6 percentage points higher than the British average. This is likely to be driven by typical wage growth in Scotland outpacing much of the rest of the UK over the 2000s. 6 6 C D Arcy and G Kelly, State of working Scotland: living standards, jobs and pay, Resolution Foundation, January

17 Figure 7: Low pay status in 2016 by region in 2006 Percentage point difference from overall average Source: Resolution Foundation analysis of ONS, NESPD The North East has the highest proportion of people stuck in the low pay trap 3.4 percentage points higher than the average across Britain. London is the other region with a particularly notable divergence from the national trends. The proportions of people escaping low pay and getting stuck are both lower than the British average. The share of cyclers on the other hand is nearly 8 percentage points higher than average. That is of particular interest given cyclers are also more common in the other higher-paying parts of the UK the South East and the East. The precise driver of this trend is unclear. One hypothesis is that because low pay is less common overall in these regions, with the going rate at the lower end of the labour market closer to the low pay threshold, it may be that more people fluctuate just above and below low pay threshold rather than remaining consistently stuck. Beyond these headline characteristics, there are of course a variety of other attributes that are important when assessing who gets stuck in low pay and who escapes. Some of these are captured in Table 1 below, which also introduces another element to the analysis: what we term dynamic characteristics. In most circumstances, the attributes considered above were consistent for low-paid individuals in 2006 all the way through to But other factors like the industry, 7 Tavle 1 shows that despite the potential for moving region to lead to progression, only a minority appear to have done so over the decade. 13

18 occupation or firm worked in, as well as whether someone works part- or full-time can and do change more frequently. And crucially, that change may tell us something important about who gets stuck in low pay and who escapes. 8 The first two rows of Table 1 shine a light on an often raised component of the low pay debate: part-time vs. full-time working. As organisations like Timewise have highlighted, the dearth of quality part-time roles often penalises those seeking to combine caring responsibilities with decently paid work. It is noteworthy that in the initial period (2006 in this instance), those who would go on to be escapers were relatively evenly divided between part-time and full-time work. But by 2016, 71 per cent of escapers were working full-time. While this may once again reflect the fact that students often work part-time in low-paid roles but go on to earn higher wages in full-time positions after graduating, the lack of well-paid part-time roles, especially for mothers with young children, is likely to be a factor. The availability of low-paying jobs may be one of the drivers behind trends visible in employer type. While 8 per cent of those who would go on to escape were employed by central government in 2006, by 2016, 17 per cent of escapers worked in central government. Such jobs tend to be generally higher-paying and can often offer more structured progression routes than in the private sector. 8 Before discussing in further detail, it should be noted that these are descriptive statistics; that is, they provide information on who gets stuck and who escapes. But to understand whether a certain characteristic like, for instance, working for a small firm, makes a meaningful difference, econometric analysis is required. 14

19 Table 1: Low paid employees in 2006 by whether they escape low pay within a decade and initial and final period characteristics Notes: * denotes a group of less than 1 per cent Source: Resolution Foundation analysis of ONS, NESPD One interpretation of central government s increased representation among escapers is that this is due to compositional changes in its workforce. The fall in the number of low-paying roles within central government over this period may have 15

20 pushed those made redundant into the private sector, thereby improving the statistics for central government. Another more positive explanation is that it is larger; people who move into roles in, for instance, the civil service may benefit from vacant positions arising more often than someone employed by a small firm. This theory, however, does not appear to be supported when we explore how change has occurred in large firms. There has been no shift towards large firms over the decade. This is not to deny that large firms can offer progression opportunities. But it may be that some large firms supermarkets and hotel chains for example represent a great deal of low-wage employment and they do not offer structured progression. Turning next to occupations, Table 1 shows, as would be expected, that escapers are more likely to have moved into generally better-paying types of jobs than those who are stuck or cycling. For example, just 1 per cent of our sample in 2006 were in professional roles. By 2016 this had grown to 11 per cent. In some senses, this is to be expected, describing as it does the fact that professional occupations are usually well-paid relative to the median. However, it is relevant given the discussion over the extent to which traditionally working-class jobs can offer a career and higher earnings over time. Figure 8 reflects this pattern too, with two-thirds (66 per cent) of those who were stuck in 2016 having remained in the same occupational group over the past decade. It is therefore the case that people who move their occupation from routine or manual to a professional-managerial are more likely to have escaped low pay. This is a question which comes through in our analysis of industries too. Noticeably, there is a shift out of the two largest low-paying sectors; together, wholesale and retail and accommodation and food services employed a combined 43 per cent of future escapers in 2006 but just 22 per cent in The opposite trend was visible in health and social work (12 per cent to 19 per cent of escapers) and education (14 per cent to 19 per cent of escapers) over the decade, with escapers more likely to have moved into these industries. This may reflect that fact that with major exceptions like care work low pay is less prevalent in these sectors than retail or hospitality, with more higher-paying positions to progress into. While this discussion of change over a decade in a person s working life offers fascinating insights into shifts over time, it overlooks an even more crucial element: how regularly a person is in work. Figure 8 identifies that escapers spent more years on average in employment over the following decade (8.3 years) than cyclers (5.9) or the stuck (4.5). While it is clear that, on average, cyclers and the stuck have not become completely detached from the labour market, the gap is a relatively large one. It may be unrealistic to expect the typical low-paid person to make significant pay progression while only working in less than half the years of the following decade. And while some in this group may face a significant challenge in finding and sustaining work without considering progression, the issue of part-time employment and the options available to mothers returning to the labour market are likely to be influential here once again. The proportion of those years spent in part-time 16

21 employment differs across the groups too. When in work, escapers spent one in three years in part-time work while for the stuck this was closer to two-thirds (63 per cent). Figure 8: Characteristics of the low paid in 2006 by their status in 2016 Source: Resolution Foundation analysis of ONS, NESPD Figure 8 draws attention to another important element of the low pay question. The long-term stuck those who had been stuck in low pay in the previous decade ( ) comprises 41 per cent of those that were still stuck in But encouragingly, it also shows that people who have already been stuck for a decade can also go on to progress, comprising 19 per cent of those who escaped in Table 2 distils some of the above-discussed attributes of each group. In general, cyclers fall somewhere between the two, for example, with a smaller share of women than within the stuck but also with men comprising a smaller share than within the escapers. The next section explores which of these characteristics are correlated with escaping from low pay. Table 2: More likely to be... than the average low-paid employee, 2006 Source: Resolution Foundation analysis of ONS, NESPD 17

22 Section 3 What helps people escape low pay? Having discussed the composition of each of our groups, we next turn to the key question: which of these characteristics appear to be linked to getting stuck or escaping? To do this, we carry out regression analysis, controlling for a range of factors to try to isolate key attributes of interest. Further discussion of the methodological approach taken can be found in the Annex. In the analysis, we once again focus on those who were low paid in 2006 and follow them up to But we also assess whether there have been shifts over time whether some factors have become more strongly (or weakly) associated with escaping low pay. We begin by assessing the difference in progression rates by age. We take as our reference group those aged in the initial period. Compared to this group of employees, every other age group is less likely to progress. While the effect is negative for those in their 20s, it is small and insignificant. But for the groups aged and upwards in 2006, the effect size becomes larger and more negative. Those in their late 30s are already 33 percentage points less likely to progress than a year old with the size of the gap growing for older groups. Older workers those in their 50s and above who are closer to retirement may stand to gain less from progression, with a few years of potentially higher pay not outweighing the difficulties associated with taking on greater responsibility. With younger employees on average more educated than their older counterparts, this may also mean older employees miss out. But given the effects are sizeable and negative from people s late 30s and onwards, this suggests progression is proving difficult for many prime-age people. While actions to improve education and training are often targeted at young people, older workers taking a broad definition of this term should not be overlooked. We next turn to gaps between men and women. As we saw in the descriptive analysis in Section 2, women comprise more of the low paid and are overrepresented among the stuck. But to what extent is this linked to the fact that they tend to, for example, work in occupations or industries that are difficult to escape from or spend fewer years in employment which acts as a brake on their progress? Our analysis finds that, as shown in Table 3, even controlling for these important factors sex still emerges as a characteristic that is linked to progressing from low pay, with women less likely to escape. 18

23 Table 3: Individual factors significantly associated with having escaped low pay in 2016 among the low paid in 2006 Source: Resolution Foundation analysis of ONS, NESPD Because of the limited personal information provided in the NESPD, we cannot comment on the role played in this trend by life events, particularly having children. But one proxy for this is the additional effect that age has on women s likelihood to progress. The interaction of age and sex is more complicated than the effect of age discussed above. Being a woman aged in the initial period was negatively and significantly linked with escaping from low pay, with a smaller and insignificant effect for women aged This is in contrast to women aged 36-55, with each five-year age band in this group more likely to escape more likely to escape low pay than their younger female colleagues. As discussed, we can t isolate the impact of family situations. However, it may be that these women in this younger age group are more likely to have children in their 20s and early 30s, thereby leaving the labour market for a period which plays a key role. By contrast, the more favourable outcomes for women in their late 30s and older may be related to the point in time at which their children begin to attend primary school, meaning a return to work, and better-paying work, is possible. Again, this question of the availability of better-paying work that fits with caring responsibilities often shouldered by women is central to the low pay debate. The previous Section highlighted that escapers were more likely to have switched from part-time employment into full-time positions over the following decade. The negative effect of part-time work appears to be supported by our regression analysis. Controlling for relevant considerations for which we have data, the higher the number of years a person spends in part-time work, the less likely they are to progress. Our analysis does however suggest that staying in employment, whether 19

24 part-time or full-time, is beneficial to escaping. The more years a person remained in employment across the following decade, the more likely they were to escape. The importance of the type of employer is also confirmed by our analysis. Working in the private sector was negatively linked with escaping from low pay while working in central government was positively linked with escaping low pay. But some of the largest effect sizes are visible when we break down the analysis by industry. Compared to all those who didn t remain working in accommodation and food services, employees who stayed in this sector were significantly less likely to escape. What Table 3 doesn t capture is change over time, that is, whether the impact of these factors has become more strongly or weakly associated with progression. Due to changes in the way occupations are coded in the data, prior to 2002 we cannot control for the role played by occupation. Over this period, there is little evidence of significant shifts. But stretching back over a longer time span to those low paid in 1996 and their status in 2016 though without occupational controls, some changes do become apparent. On age, we find that the size of the negative effect for those in their 20s has dwindled over time. While they still remain less likely to escape than year olds, there is no longer a significant relationship. Turning next to the kind of organisation worked for, while working for a private company has always been negatively and significantly linked to escaping from low pay, the size of the effect has become slightly more negative over time. Again, this could be linked to the reduction in low-paying jobs in central government over time meaning that the private sector has not necessarily become worse for low-paid individuals already working there but that a greater proportion of the low paid are in the private sector now. There is evidence of a changing role among industries. Remaining in manufacturing, mining and quarrying has become more negatively linked to escaping low pay over time. A relevant consideration may be that as the industry has shrunk in recent decades, this may limit the number of positions that people can move up into. A similar trend of a growing negative effect is visible in health and social work. After moving from being negatively linked to progression among the late 1990s cohorts to becoming positive (though insignificant) in the early 2000s, recent years have seen the industry become negatively associated to progression once again. The opposite trend has occurred in wholesale and retail. Though it is negatively linked to escaping low pay (when not controlling for occupations), the strength of the negative effect has weakened over time. This suggests that while it is still a difficult sector to escape from low pay in, the size of the negative impact has lessened. A similar pattern applies in education, which has changed from being negatively and 20

25 significantly associated with escaping low pay to its current position of being slightly positively linked to progression, though this finding is not significant. The next section of this report the conclusion summarises our findings before suggesting some possible policy responses that employers and government could take to tackle the issues the analysis has raised. 21

26 Section 4 Conclusion Our analysis confirms that, for most people, low pay is not acting as a first rung on the pay ladder. Just one in six employees made sustained progress out of low-wage work in the most recent period. This represents a slight improvement over recent decades but is far from the transformational shift required. To achieve meaningful progress, a more comprehensive approach will be needed. Employers clearly have a role to play, particularly in sectors like hospitality and retail in which low pay is more prevalent and progression appears limited. Efforts to make these industries attractive are likely to be all the more crucial in coming years given existing skills gaps that could be exacerbated by a potential fall in migration from the EU. Some firms in these sectors, often working with organisations like the Living Wage Foundation and the British Retail Consortium, have taken the lead, recognising the importance of career routes and boosting the supply of better-paying part-time positions. This last point the need for higher-quality part-time work is especially vital for women, who our research has identified as in need of greater support. The lack of decently-paid jobs offered on a part-time or flexible basis acts as an insurmountable barrier for some women. But employers are unlikely to solve these problems alone. The government s upcoming industrial strategy offers the chance to work with employers in low-paying sectors to design jobs that offers progression opportunities to a wider pool of employees, but that first and foremost moves away from a business model that relies above all on very cheap labour. The introduction of Universal Credit provides another chance to improve work incentives for second earners, thereby making pay progression more attractive. 9 While second earner incentives are a longstanding issue in the benefits system, the new regime brings innovation too. The data available in Universal Credit could be used to identify long-term low-paid individuals and signpost them to progression support. The exact form this support should take could be tested, with Metro Mayors, local authorities and LEPs all with a role to play in matching the approach to local needs and conditions. But alongside this locally driven action, the creation of a national body that could oversee these new progression trials, ensuring they are high quality and helping to spread best practice would illustrate that movement out of low pay is being taken seriously. Another Universal Credit innovation is in-work conditionality. With exceptions for those with young children, caring responsibilities or ill health, people earning less than the equivalent of 35 hours a week at the wage floor will be required to either increase the number of hours they work or how much they earn per hour. While 9 For further discussion of policy responses that could boost pay progression, in particular the role played by Universal Credit, see D Finch, Moving on up: Enabling earnings progression in the UK labour market in ed. S Clarke, Work in Brexit Britain: Reshaping the Nation s Labour Market, Resolution Foundation, July

27 many employees may feel the former is the most straightforward option, the latter could potentially see more people especially low-paid mothers in search of higher-paid positions. Taking this as a trigger, additional support could be offered to help kickstart progression among older workers, another group for whom escaping low pay appears difficult. Poor progression out of low-paying work won t be solved overnight. But boosting the number of people that escape from low pay, and spreading opportunities to groups with lower progression rates, should be a focus for government and employers alike in the years to come. 23

28 Annex To identify the key factors associated with progression out of low pay, we ran a probit regression where the dependent variable is a binary variable equal to one if an individual was an escaper, which is regressed against the available individual and employment characteristics variables. The estimation is restricted to just those employees who were low paid in The model therefore identifies the factors associated with escaping low pay, conditional on being low paid in initial period and controlling for observable characteristics. The table below presents the regression results. Unfortunately due to data limitations we are unable to assess how education and skills and training influence pay progression, since this information is not covered in the NESPD. It would seem likely that having a low level of education and skills may hinder an individual s opportunities for pay progression. Similarly, it would be very useful to know whether training can support employees to progress, and what types of training are most effective. As a robustness check, but also to understand how the factors associated with escaping low pay have changed over time, this analysis was repeated separately for the years back to The occupation group dummies were omitted from the models because these were not consistent between 2006 and prior years though we do include occupation dummies as far back as A summary of the results is presented below. 24

29 Table 4: Probit regression results: probability of escaping low pay by 2016 conditional on being low paid in initial year (2006) Employee characteristics Male *** Initial year characteristics: Low paid in 2006 and have only ever held low paid jobs in the previous decade *** Age group (reference category: 16-20) 21 to to to ** 36 to *** 41 to *** 46 to *** 51 to *** 56 to *** 61 to *** *** Additional age effect for females 21 to *** 26 to to * 36 to *** 41 to ** 46 to *** 51 to *** 56 to to Dynamic characteristics (subsequent decade): Average number of years in subsequent decade ( ) in employment *** in part-time employment *** switched jobs within last 12 months Average number of years in subsequent decade employed by... very small firm (<10 employees) very large firm (1000+ employees) ** private company *** central government *** local authority ** Remained working in the same industry (reference category: all those not remaining in these industries) Manufacturing, mining & quarrying ** Wholesale and retail ** Accommodation and food services *** Education Health & social work ** Sample size 26,733 Pseudo-Rsq Notes: Robust standard errors in parentheses.*** p<0.01, ** p<0.05, * p<0.1 Source: Resolution Foundation analysis of ONS, NESPD 25

30 1

31 Crown copyright 2017 Any enquiries regarding this publication should be sent to us at the Social Mobility Commission, Sanctuary Buildings, 20 Great Smith Street, London, SW1P 3BT. This document is also available from our website at: